Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
    pvanheusden
    @pvanheusden:matrix.org
    [m]
    well this is a rather old Galaxy server... upgraded over the years, but obviously not the conda part
    I see the latest-on-offer is 4.10.0
    pvanheusden
    @pvanheusden:matrix.org
    [m]
    upgrading conda seems to have fixed the problem
    Björn Grüning
    @bgruening
    :+1:
    pvanheusden
    @pvanheusden:matrix.org
    [m]
    what is the option to re-use computed datasets if the inputs stay the same? I'd like to try this to see if it can save some time with IRIDA (which is bad about caching computed results)
    I recall mvdbeek working on this some time back
    Helena Rasche
    @hexylena:matrix.org
    [m]
    it's in user preferences (may have to copy from sample?), it is not very aggressive in the caching, was my experience. limited to within a user, i think precisely the same HID even in a history? I'm not sure about all of the restrictions
    pvanheusden
    @pvanheusden:matrix.org
    [m]
    ah ok...
    so IRIDA runs everything as a single user.
    but I don't know if there is a way to say "always try and re-use"
    Helena Rasche
    @hexylena:matrix.org
    [m]
    eu has it enabled if you want to play around with it.
    pvanheusden
    @pvanheusden:matrix.org
    [m]
    I fear that because of how IRIDA loads things it will probably not be used... need to poke some
    Martin
    @martinwolst
    Does anyone have an estimate of how large the Galaxy postgres database is likely to get after ~1 year of running by 15-20 users? (I realise it depends what the users do, but I need a worst case estimate for backups etc.) Thanks!
    Pablo Moreno
    @pcm32
    I’m having trouble with our resubmissions on LSF when moving from 19.05 to 20.09. I’m trying to compare against other runners, but I’m not sure which ones support resubmissions. Do you use Galaxy in an environment where job automatic resubmissions (ie. when OOM) work? Could you please mention it here? Thanks!
    8 replies
    dickgroenenberg
    @dickgroenenberg
    I posted this in the 'Galaxy Training Network' yesterday, but this might be more of an Admin topic: Just finished the "Galaxy installation with Ansible" tutorial. Up to the NGINX paragraph everything works fine (that is running the playbook after Systemd allows access to Galaxy over port: 8080). However after installing NGINX and running the playbook, systemctl shows galaxy and nginx.service are running, but the Galaxy webpage cannot be opened. The ngnix error.log says:"/srv/nginx/_well-known_root/.well-known/security.txt" failed (2: No such file or directory)". Any suggestions on how to solve this issue?
    Helena Rasche
    @hexylena:matrix.org
    [m]
    That error log probably isn't the real failure
    sorry, saw the GTN post but got too busy to follow up
    can you share more of your error/access logs?
    what does the webpage show? which error code?
    dickgroenenberg
    @dickgroenenberg
    Webpage only shows "can't open page "domainname:80" because Safari can't establish a secure connection to the server
    Helena Rasche
    @hexylena:matrix.org
    [m]
    oh this sounds like a safari problem
    try https://domainname in any other browser
    you should see a warning for an untrusted certificate
    some browsers think they know better, and refuse to give you a "yes I know what I'm doing" button
    dickgroenenberg
    @dickgroenenberg
    chrome says "your connection to this site is not secure" and then shows "site can't be reached"
    Helena Rasche
    @hexylena:matrix.org
    [m]
    is it saying that for :80 or :443
    and: the first "not secure" is expected, right?
    you've setup a letsencrypt staging certificate
    so that's completely normal
    after that a "site can't be reached" is more interesting and maybe where the next step for you is
    dickgroenenberg
    @dickgroenenberg
    neither :80 nor :443 make a connection. So changing 'staging' to 'production' should do the trick?
    Helena Rasche
    @hexylena:matrix.org
    [m]
    no, that shouldn't help
    if it isn't working with staging you're going to have the same trouble with production, except your certificate will be valid, it won't fix the underlying issue
    1. you've got systemctl status nginx galaxy showing both working, right?
    2. what does your nginx conf look like?
    3. what does your galaxy uwsgi conf look like?
    4. what does 'cannot make a connection' mean? Can you e.g. curl -k https://domainname? what does that return?
    5. do you have a firewall?
    dickgroenenberg
    @dickgroenenberg
    1. both are active (running)
    4.curl: (6) Could not resolve host: https
    1. no firewall, ports 80 and 8080 are open on OpenStack
    Helena Rasche
    @hexylena:matrix.org
    [m]
    you need port 443 too
    (for the full admin training we also open 5671 for AMQP, in case you plan to go further in the training)
    that explains the connection issues :)
    dickgroenenberg
    @dickgroenenberg
    I'll check it out. Thanks for the support Helena!
    Helena Rasche
    @hexylena:matrix.org
    [m]
    sure thing, we're always happy to see more folks using it!
    we can add a note to the training material about which ports you need open, that's a good catch.
    we prepare the training machines, but didn't make this explicit for people using their own ones
    dickgroenenberg
    @dickgroenenberg
    Indeed it was port 443 !
    Helena Rasche
    @hexylena:matrix.org
    [m]
    awesome! fantastic
    success with the rest!
    martenson
    @martenson:matrix.org
    [m]
    When this operation (https://github.com/galaxyproject/galaxy/blob/dev/lib/galaxy/objectstore/s3.py#L623) fails - e.g. when the cache has already been emptied, the actual remote object is not deleted. Is this causing Galaxy to leave orphaned files behind?
    Because you can purge and delete histories with such datasets, however they remain in the remote objectstore.
    Curtis Ross
    @cross12tamu

    Hello on a Saturday :grimacing: ,

    I'm having some issues with some jobs not getting passed queue in condor. I could not quite follow what was going on in the galaxy logs, but thought it was just some job not properly showing that were perhaps hung. I ended my jobs that were all in the queue. I still had the same error "looping" through in the logs:
    So sorry for the poke on a Saturday y'all.

    galaxy.workflow.run DEBUG 2021-04-17 20:33:58,772 [p:778116,w:0,m:2] [WorkflowRequestMonitor.monitor_thread] Marking step 281170 outputs of invocation 14208 delayed (dependent collection [1437] not yet populated with datasets)
    galaxy.workflow.run DEBUG 2021-04-17 20:33:58,772 [p:778116,w:0,m:2] [WorkflowRequestMonitor.monitor_thread] Marking step 281171 outputs of invocation 14208 delayed (dependent step [281170] delayed, so this step must be delayed)
    galaxy.workflow.run DEBUG 2021-04-17 20:33:58,772 [p:778116,w:0,m:2] [WorkflowRequestMonitor.monitor_thread] Marking step 281172 outputs of invocation 14208 delayed (dependent step [281171] delayed, so this step must be delayed)
    galaxy.workflow.run DEBUG 2021-04-17 20:33:58,773 [p:778116,w:0,m:2] [WorkflowRequestMonitor.monitor_thread] Marking step 281173 outputs of invocation 14208 delayed (dependent step [281172] delayed, so this step must be delayed)
    galaxy.workflow.scheduling_manager DEBUG 2021-04-17 20:33:58,778 [p:778116,w:0,m:2] [WorkflowRequestMonitor.monitor_thread] Workflow invocation [14208] scheduled
    galaxy.workflow.scheduling_manager DEBUG 2021-04-17 20:33:58,825 [p:778119,w:0,m:3] [WorkflowRequestMonitor.monitor_thread] Attempting to schedule workflow invocation [(14192,)]
    galaxy.workflow.run DEBUG 2021-04-17 20:33:59,090 [p:778119,w:0,m:3] [WorkflowRequestMonitor.monitor_thread] Marking step 282238 outputs of invocation 14192 delayed (dependent collection [1428] not yet populated with datasets)
    galaxy.workflow.run DEBUG 2021-04-17 20:33:59,090 [p:778119,w:0,m:3] [WorkflowRequestMonitor.monitor_thread] Marking step 282239 outputs of invocation 14192 delayed (dependent step [282238] delayed, so this step must be delayed)
    galaxy.workflow.run DEBUG 2021-04-17 20:33:59,090 [p:778119,w:0,m:3] [WorkflowRequestMonitor.monitor_thread] Marking step 282240 outputs of invocation 14192 delayed (dependent step [282239] delayed, so this step must be delayed)
    galaxy.workflow.run DEBUG 2021-04-17 20:33:59,091 [p:778119,w:0,m:3] [WorkflowRequestMonitor.monitor_thread] Marking step 282241 outputs of invocation 14192 delayed (dependent step [282239] delayed, so this step must be delayed)
    galaxy.workflow.run DEBUG 2021-04-17 20:33:59,091 [p:778119,w:0,m:3] [WorkflowRequestMonitor.monitor_thread] Marking step 282242 outputs of invocation 14192 delayed (dependent step [282241] delayed, so this step must be delayed)
    ........ more ......

    And my condor_q (just sanity checking):

     ID      OWNER            SUBMITTED     RUN_TIME ST PRI SIZE CMD
    
    0 jobs; 0 completed, 0 removed, 0 idle, 0 running, 0 held, 0 suspended

    Does anyone have some sage wisdom for me?

    1 reply